DE102019210405A1 - Method for the investigation of substances, computer program product for use in this method, and mobile electronic device - Google Patents

Abstract

The present invention relates to a method for examining substances. In the method, a mobile electronic device, which has a camera with a camera lens, and a test pattern are provided. The substance to be examined (or a mixture of the substance to be examined with at least one auxiliary substance) is brought into contact with the camera lens or with a protective element arranged on or in front of the camera lens, so that a film of the substance to be examined or the mixture forms on the Forms camera lens or the protective element. With the help of the camera, an image of the test pattern is recorded while the film of the substance to be examined or the mixture is on the camera lens or the protective element, the recorded image of the test pattern showing an at least partially distorted representation of the test pattern, with at least one Part of texture features and / or image features of the at least partially distorted representation of the test pattern differs from the corresponding texture features and / or image features of the original test pattern. After at least some of the differing texture features and / or image features have been extracted from the recorded image of the test pattern, the extracted texture features and / or image features are evaluated.

Description

The present invention relates to a method for examining substances. In the method, a mobile electronic device, which has a camera with a camera lens, and a test pattern are provided. The substance to be examined (or a mixture of the substance to be examined with at least one auxiliary substance) is brought into contact with the camera lens or with a protective element arranged on or in front of the camera lens, so that a film of the substance to be examined or the mixture forms on the Forms camera lens or the protective element. With the help of the camera, an image of the test pattern is recorded while the film of the substance to be examined or the mixture is on the camera lens or the protective element, the recorded image of the test pattern showing an at least partially distorted representation of the test pattern, with at least one Part of texture features and / or image features of the at least partially distorted representation of the test pattern differs from the corresponding texture features and / or image features of the original test pattern. After at least some of the differing texture features and / or image features have been extracted from the recorded image of the test pattern, the extracted texture features and / or image features are evaluated.

Typically, qualitative and quantitative analyzes of unknown chemical substances are carried out using suitable wet chemical methods. These are usually not real-time capable, associated with high investment and operating costs and not easily accessible to a consumer. Often part or all of the available sample is used up.

In addition, there are spectro-optical methods which, on the basis of spectroscopy, enable such analyzes in real time and non-invasively (without consuming the sample). However, special devices are also required here, which are also associated with high investment and operating costs and are likewise not easily accessible to a consumer.

The use of smartphones beyond pure communication applications for use as a measuring device / tester makes the determination of ingredients cheaper and generally suitable for consumers. In this context, there are already spectro-optical methods, which, however, severely restrict broad use by consumers due to the requirement for external optical, mechanical or electronic components. In the WO 2018/115346 the requirement for external hardware is circumvented. However, spectro-optical methods in the form of reflection measurement are fundamentally unsuitable for liquids due to the phase transition of media with different optical densities. In addition, a suitable calibrated, broadband light source and a spectrally known and constant reflector are required.

On the basis of this, the object of the present invention was to provide a method for examining substances with which in particular liquid or gel-like substances can be examined in the simplest and most cost-effective way possible.

This object is achieved with regard to a method with the features of patent claim 1, with regard to a computer program product with the features of patent claim 16 and with regard to a mobile electronic device with the features of patent claim 17.

1. i) ein mobiles elektronisches Gerät bereitgestellt wird, welches eine Kamera mit einer Kameralinse aufweist,
2. ii) ein Testmuster bereitgestellt wird,
3. iii) eine zu untersuchende Substanz oder eine Mischung einer zu untersuchenden Substanz und mindestens eines Hilfsstoffs mit der Kameralinse oder mit einem auf oder vor der Kameralinse angeordneten Schutzelement in Kontakt gebracht wird, so dass sich ein Film der zu untersuchenden Substanz bzw. der Mischung auf der Kameralinse bzw. dem Schutzelement bildet,
4. iv) mindestens ein Bild des Testmusters mithilfe der Kamera aufgenommen wird, während sich der Film der zu untersuchenden Substanz bzw. der Mischung auf der Kameralinse bzw. dem Schutzelement befindet, wobei das aufgenommene Bild des Testmusters eine zumindest teilweise verzerrte Darstellung des Testmusters zeigt, wobei sich zumindest ein Teil von Texturmerkmalen und/oder Bildmerkmalen der zumindest teilweise verzerrten Darstellung des Testmusters von den entsprechenden Texturmerkmalen und/oder Bildmerkmalen des ursprünglichen Testmusters unterscheidet,
5. v) zumindest ein Teil der sich unterscheidenden Texturmerkmale und/oder Bildmerkmale aus dem aufgenommenen Bild des Testmusters extrahiert wird, und
6. vi) eine Auswertung der extrahierten Texturmerkmale und/oder Bildmerkmale erfolgt.

According to the invention, a method for examining substances is thus specified in which

1. i) a mobile electronic device is provided which has a camera with a camera lens,
2. ii) a test pattern is provided,
3. iii) a substance to be investigated or a mixture of a substance to be investigated and at least one auxiliary substance is brought into contact with the camera lens or with a protective element arranged on or in front of the camera lens, so that a film of the substance to be investigated or the mixture is on the Camera lens or protective element,
4. iv) at least one image of the test pattern is recorded with the aid of the camera while the film of the substance to be examined or the mixture is on the camera lens or the protective element, the recorded image of the test pattern showing an at least partially distorted representation of the test pattern, with at least some of the texture features and / or image features of the at least partially distorted representation of the test pattern differ from the corresponding texture features and / or image features of the original test pattern,
5. v) at least some of the differing texture features and / or image features are extracted from the recorded image of the test pattern, and
6. vi) the extracted texture features and / or image features are evaluated.

With the method according to the invention, in particular (at least partially) liquid and / or (at least partially) gel-like substances can be examined. Both artificially produced substances (e.g. motor oil, cosmetic creams, lemonade) and natural substances (e.g. blood, sebum, sweat, vegetable oil, tree resin, fruit juice) can be examined. In principle, however, the method is also suitable for examining solid substances. These can for example be mixed with a solvent as an auxiliary substance before step ii) so that a mixture is formed which can be applied as a film to the camera lens or the protective element.

In step i), a mobile electronic device is first provided which has a camera with a camera lens. The camera of the mobile electronic device is suitable for taking pictures. The camera of the mobile electronic device is preferably integrated into a housing of the mobile electronic device. The mobile electronic device preferably has a screen for displaying images. The screen can be integrated into the housing of the mobile electronic device. The mobile electronic device can be, for example, a smartphone or a tablet computer. It is also possible to use a mobile electronic camera device, e.g. a digital camera, as a mobile electronic device.

In step ii) a test pattern is provided. The test pattern is provided in such a way that an image or photo of the test pattern can be recorded with the aid of the camera of the mobile electronic device. In principle, the test pattern can be provided in any form, e.g. as a representation on a screen or as a representation or imprint on a surface, such as a sheet of paper.

The test pattern is preferably matched to the examination of the substance to be examined. The test pattern can, for example, be matched to the examination of the substance to be examined in that the test pattern is adapted to the substance to be examined itself or to at least one (e.g. known and / or previously determined) property of the substance to be examined. Additionally or alternatively, the test pattern can also be matched to the extraction performed in step v), the extraction method or a partial step of this extraction, and / or to the evaluation performed in step vi), the method of this evaluation, or a partial step of this evaluation be.

In step iii) a substance to be examined (or a mixture of a substance to be examined with an auxiliary substance, e.g. with a solvent) is brought into contact with the camera lens or with a protective element arranged on or in front of the camera lens. As a result, a certain amount of the substance (or the mixture) to be examined can remain in the form of a film on the camera lens or the protective element. The camera lens or the protective element can thus be covered at least partially (or also completely) with the film of the substance to be examined (or the mixture). As a result, the light rays that enter the camera lens when a picture or photo is taken with the camera first pass through the film of the substance to be examined (or the mixture) before they reach the camera lens. Here, the light rays falling into the camera lens are influenced by the substance to be examined. The substance to be examined thus acts as a kind of optical filter.

The film of the substance to be examined (or the mixture) formed in step iii) preferably has a transparency of at least 10%, preferably at least 30%, particularly preferably at least 50%, very particularly preferably at least 70%, for electromagnetic radiation of one or more several wavelengths in the wavelength range from 1 nm to 100,000 nm, preferably electromagnetic radiation in one or more wavelengths in the wavelength range from 10 nm to 10,000 nm, particularly preferably for electromagnetic radiation in one or more wavelengths in the visible wavelength range, in the near UV range and in the IR range , on.

The protective element is understood to mean an element which is attached to the camera lens or in front of the camera lens in order to protect the camera lens from external influences. This can be, for example, a protective glass, a protective pane or a protective film. The protective element is preferably at least partially permeable or transparent to electromagnetic radiation of one or more wavelengths in the wavelength range from 1 nm to 100,000 nm, preferably electromagnetic radiation of one or more wavelengths in the wavelength range from 10 nm to 10,000 nm, particularly preferably for electromagnetic radiation one or more Wavelengths in the visible wavelength range, in the near UV range and in the IR range. Said permeability or transparency can be, for example, at least 50%, preferably at least 70%, particularly preferably at least 90%.

The substance to be examined (or the mixture with the at least one auxiliary substance) can preferably be brought into contact with the camera lens or the protective element in that the camera lens or the protective element approaches the substance to be examined (or the mixture) is applied or the substance to be examined (or the mixture) is transferred to the camera lens or the protective element with a cloth.

In step iv), at least one image or photo of the test pattern is then recorded with the aid of the camera of the mobile electronic device while the film of the substance to be examined (or the mixture) is on the camera lens or the protective element. The light rays entering the camera lens when the image or photo is taken first pass the film of the substance to be examined (or the mixture) (located on the camera lens or the protective element) before they reach the camera lens and are thereby influenced by the substance to be examined, for example at least partially deflected, for example by refraction and / or diffraction (diffraction) on the substance to be examined. The substance to be examined ultimately acts as a type of optical filter. As a result of this influencing, the recorded image or photo of the test pattern shows an at least partially distorted representation of the original (i.e. originally provided in step ii)) test pattern. Here, at least some of the texture features and / or image features of the at least partially distorted representation of the test pattern differ from the corresponding texture features and / or image features of the original (i.e. originally provided in step ii)) test pattern. This means that the texture features and / or image features of the representation of the pattern shown on the recorded image are partially shifted and / or look different and / or are arranged differently to one another than the corresponding texture features and / or image features of the original test pattern.

Texture features and / or image features can, for example, be certain elements of the pattern, such as the course and / or length of certain lines of the pattern; Course and / or length of certain borders or edges of certain areas or sections of the pattern; Distance and / or position of certain points of the pattern to one another or to certain lines or areas of the pattern; Shape of certain areas of the pattern; Intersections of certain lines of the pattern; Symmetry of the pattern or certain sections of the pattern; Course and / or position of axes of symmetry of the pattern; Position of symmetry points of the pattern.

In step v) at least some of the differing texture features and / or image features - ie the texture features and / or image features of the at least partially distorted representation of the test pattern that differ from the corresponding texture features and / or image features of the original test pattern - are removed from the extracted image of the test pattern. This extraction can take place, for example, with the aid of software, for example a computer program product installed on the mobile electronic device. Before the extraction, the differing texture features and / or image features can first be determined or identified, for example by comparing the recorded image with the original test pattern (or with an additionally recorded reference image). This determination or identification can also take place, for example, using the software mentioned. Such features are preferably extracted which change significantly as a function of the influence caused by the substance to be examined. As evidence, the change in the distance of feature histograms is preferably calculated (between measurement of a reference - i.e. without substance - and measurement with substance) or, in the case of point illumination, the point spread function. A significant change occurs, for example, when artifacts (e.g. displacements of points) occur in certain frequency components of the recorded image, which are caused precisely by the optical properties of the substance to be measured as an optical filter. If only ambient light is used for the measurement, which is typically quite homogeneous, there are a large number of pixels that can be aggregated via a histogram. This can be distinguished from the case of point lighting (e.g. by flash light), in which a point spread function can be used.

Particularly preferred image filters / features are the class of texture features. If several texture features are preferably used, these can be compiled as an aggregated feature vector and evaluated by the following algorithm. As a supplement, e.g. further feature filters and color histograms can be used explicitly or through filter masks that are implicitly obtained in the context of machine learning, especially deep learning, e.g. as layers in convolutional neural networks.

The extraction of the texture features taking place in step v) is preferably and / or Image features matched to the investigation of the substance to be investigated. For example, texture features and / or image features are extracted that allow a (particularly precise) conclusion about the substance to be examined with regard to a specific application in the subsequent evaluation. It is therefore preferred that the texture features and / or image features extracted in step v) are matched to the examination of the substance to be examined.

Finally, in step vi) the texture features and / or image features extracted in step v) are evaluated. Here, based on the extracted texture features and / or image features, conclusions can be drawn about the substance to be examined (e.g. its quality, one or more of its properties, its belonging to a certain type of substance or substance class, its authenticity and / or its chemical composition ) to be hit. In the evaluation, for example, the change in certain texture features and / or image features from the original test pattern (or a reference image) to the image recorded in the presence of the substance or the mixture can be determined. For example, the change in a length or a course of a specific line or the change in the position of a specific point or two specific points relative to one another could be determined. For example, feature vectors can also be determined here. From the specific change in the texture features and / or image features, the above-mentioned conclusions can then be made about the substance to be examined, e.g. by comparison with specific values from a database or by comparison with values determined in previous measurements. Instead of relative values, absolute values can also be determined (e.g. without using a reference), which are then compared with specific values from a database or by comparison with values determined in previous measurements. The evaluation in step vi) can take place, for example, with the aid of software, e.g. a computer program m product installed on the mobile electronic device.

The evaluation carried out in step vi) is preferably matched to the examination of the substance to be examined.

• - das in Schritt ii) bereitgestellte Testmuster, und/oder
• - die in Schritt v) extrahierten Texturmerkmale und/oder Bildmerkmale, und/oder
• - die in Schritt vi) durchgeführte Auswertung

auf die Untersuchung der zu untersuchenden Substanz abgestimmt. Diese Abstimmung kann beispielsweise durch eine anwendungsspezifische Optimierung erfolgen. Anwendungsspezifisch bedeutet hierbei, dass eine Anpassung der genannten Verfahrensmerkmale an den Zweck der Untersuchung vorgenommen werden kann. Soll beispielsweise die Qualität von Motoröl untersucht werden, können die genannten Merkmale - d.h. das bereitgestellte Testmuster und/oder das Extrahieren in Schritt v) und/oder die Auswertung in Schritt vi) - so angepasst werden, dass sie besonders für die Untersuchung von Motoröl und für die Bestimmung von dessen Qualität geeignet sind. Zu einer solchen Optimierung kann das erfindungsgemäße Verfahren selbst eingesetzt werden, indem es beispielsweise mehrfach mit verschiedenen bekannten Motorölen durchgeführt wird und dabei durch bestimmte während der Auswertung erfolgende Optimierungsschritte die besonders geeigneten Verfahrensmerkmale herausgefunden werden. Die Bestimmung der Qualität von Motoröl ist hierbei lediglich ein Beispiel. Andere Beispiele wären z.B. die Bestimmung des Fettanteils in Hauttalg oder Haartalg und die Bestimmung des Zuckergehalts in Lebensmitteln bzw. Getränken.In a very particularly preferred variant of the method according to the invention are

• - the test pattern provided in step ii), and / or
• the texture features and / or image features extracted in step v), and / or
• - The evaluation carried out in step vi)

matched to the investigation of the substance to be investigated. This coordination can take place, for example, through an application-specific optimization. Application-specific here means that the stated procedural features can be adapted to the purpose of the investigation. If, for example, the quality of engine oil is to be examined, the features mentioned - ie the provided test pattern and / or the extraction in step v) and / or the evaluation in step vi) - can be adapted so that they are particularly useful for the examination of engine oil and are suitable for determining its quality. The method according to the invention itself can be used for such an optimization, for example by performing it several times with various known engine oils and thereby finding out the particularly suitable method features through certain optimization steps that take place during the evaluation. Determining the quality of engine oil is just one example. Other examples would be the determination of the fat content in sebum or hair sebum and the determination of the sugar content in food or beverages.

The coordination or optimization of the features mentioned - ie the test pattern provided and / or the texture features / image features extracted in step v) and / or the evaluation in step vi) - to the examination of the substance to be examined can take place by means of mathematical modeling which preferably several steps are run through iteratively. Particularly preferably, one or more measurements are first carried out with a substance with regard to a desired application on one or more test patterns, then a plurality of test patterns suitable for the application are selected, then several texture features and / or image features of the test pattern (s) with regard to their suitability for use investigated in the desired application and selected suitable ones, then combinations of certain (previously selected) test patterns and texture features / image features are determined, their use in a measurement of the substance and subsequent evaluation enable a significant statement regarding the desired application, the evaluation steps used here to the specific combination of test pattern and texture features / image features can be adapted or optimized in order to obtain a result that is as meaningful as possible, and finally a determination of one for the desired one Application particularly suitable specific combination of test pattern, extracted texture features / image features and evaluation or evaluation steps carried out.

1. 1. Erfassung von Messdaten in einer Problemdomain per systematischem Experiment oder systematischen Experimenten,
2. 2. Festlegung möglicher Testmuster (z.B. Unterscheidung, ob Druck oder leuchtendes Muster, je nachdem was in der Applikation realisiert werden kann),
3. 3. Systematische Anwendung von Bildfiltern (d.h. Texturmerkmalen und/oder Bildmerkmalen) sowie Vergleich mit Referenzbild oder (z.B. intern auf dem Gerät abgespeicherter) ursprünglicher Darstellung des Musters,
4. 4. Anlernen eines Detektionssystems auf vorhandenen Kombinationen von Testmuster und Bildfiltern (d.h. Texturmerkmalen und/oder Bildmerkmalen),
5. 5. Ermittlung einer entscheidungsrelevanten Kombination aus Bildfilter (d.h. Texturmerkmalen und/oder Bildmerkmalen) und Testmuster,
6. 6. Festlegung des Testmusters und der Bildfilter (d.h. Texturmerkmale und/oder Bildmerkmale) für die konkrete Realisierung der Technologie für die Problemdomaine,
7. 7. optional: Optimierung und weitere Validierung an erweiterter Datenmenge

For example, when coordinating or optimizing the features mentioned - ie the test pattern provided and / or the texture features / image features extracted in step v) and / or the evaluation in step vi) - the following steps can be carried out:

1. 1. Acquisition of measurement data in a problem domain by means of systematic experiment or systematic experiments,
2. 2. Definition of possible test patterns (e.g. differentiation, whether print or luminous pattern, depending on what can be implemented in the application),
3. 3. Systematic application of image filters (i.e. texture features and / or image features) as well as comparison with reference image or (e.g. internally stored on the device) original representation of the pattern,
4. 4. Learning of a detection system on existing combinations of test patterns and image filters (i.e. texture features and / or image features),
5. 5. Determination of a decision-relevant combination of image filter (i.e. texture features and / or image features) and test pattern,
6. 6. Determination of the test pattern and the image filters (i.e. texture features and / or image features) for the concrete implementation of the technology for the problem domain,
7. 7. optional: Optimization and further validation on an extended amount of data

Points 3, 4 and 5 can be iteratively run through several times and represent the actual optimization of the test pattern and image filter (ie texture features and / or image features) against the background of the specific application. Such an optimization can also be performed within the evaluation in step vi) of the method according to the invention take place. For example, the extracted structural features and / or image features can be optimized for the test pattern and / or the substance to be examined.

1. 1) Bildoperation zur Bewertung der globalen Bildentropie,
2. 2) Bildoperation zur Bewertung der globalen Weichzeichnung (eng. Bluriness) im Bild,
3. 3) Bildoperation zur Bewertung des Mittelwertes, der Standardabweichung und des Kontrastes eines lokalen Bildausschnittes,
4. 4) Bildoperation zur Bewertung der Orientierung und Stärke von Bildkanten (schwarz-weiß Übergänge) mittels, z.B. Gaborfilter,
5. 5) Bildoperation zur Errechnung einer 2D Fourier Transformation zur Bewertung von periodischen Mustern im Bild.

For example, one or more of the following image filter operations can be used to detect features:

1. 1) image operation to evaluate the global image entropy,
2. 2) image operation to evaluate the global bluriness in the image,
3. 3) Image operation to evaluate the mean value, standard deviation and contrast of a local image section,
4. 4) Image operation to evaluate the orientation and strength of image edges (black-white transitions) by means of, e.g. Gabor filters,
5. 5) Image operation to calculate a 2D Fourier transformation for evaluating periodic patterns in the image.

Each image in the problem-specific database can be filtered with all or some of the image operators mentioned. For the ultimate feature extraction, the image features obtained can be used as absolute values, as relative values for an on-site recording of the test pattern with a clean lens, as relative values for an internally stored recording of the test pattern with a clean lens. Possibilities for a relative comparison are, for example, the calculation of the pixel-by-pixel deviation from the measurement image to the reference image using the mean square error (MSER) or the calculation of the distance of the histograms, formed on the measurement image and the reference image (e.g. Kullback-Leibler divergence, Chi-Squared divergence ).

For example, the parameters of a mathematical model can be adapted by means of a learning algorithm in such a way that the error in the prediction of the class based on the feature vector becomes minimal. Several methods, e.g. Artificial Neural Networks (ANN), Support Vector Machines (SVM) or Partial Least Squares (PLS) models can be tested for their suitability and selected based on a performance criterion. In order to find the optimal combination of test pattern and feature extraction, a problem-specific ordering of the importance of the features can be carried out using the internal weighting of the features in the mathematical model. The model creation can then be repeated systematically, with successive features being added to the input of the model in order of their importance. This process can be carried out until an equivalent performance compared to the system based on all features is achieved. This consideration then results in an application-specific combination of feature and test pattern for the respective task.

With the method according to the invention, on the one hand, substances can be examined, on the other hand, the examination of substances, ie the method itself, can also be optimized with regard to a specific application. For example, the method can first be carried out with one known substance or several known substances. Here, within the evaluation, the method can be optimized for the known substance (s) (or, for example, for determining a certain property of the substances or the quality of the substance). With this optimization, for example, a certain test pattern can be determined which is particularly suitable for the desired examination of this substance (s), and / or certain texture features and / or image features can be determined which are particularly suitable for the desired examination of this substance (s). suitable. In addition, certain changes in texture features and / or image features can also be determined in this optimization, which indicate certain properties of the substance or allow other conclusions to be drawn about the substance. The method optimized in this way can then be used to detect the substance (s) or to examine the substance (s), for example with regard to their quality or certain properties.

In the method according to the invention, the substance to be examined is not analyzed spectrally, but based on its effect as an optical filter, i.e. based on the refraction / diffraction (diffraction) caused by the substance to be examined, which can be measured as transmission. The physical measuring principle is thus geometrical and not spectral, and therefore largely independent of difficult-to-control properties of the lighting and the ambient light. The measuring principle is based on fundamental physical discoveries, including the wave optics formulated by Christiaan Huygens around 1650, the Huygens-Fresnel principle, the Kirchhoff diffraction integral with its two borderline cases of Fresnel diffraction (diverging point radiation source) and Fraunhofer diffraction (parallel light rays as radiation source) .

With the method according to the invention, however, an application-suitable measurement method is provided which goes beyond the pure demonstration of a physical measurement effect. Thus, not only is a simple image comparison made between the original representation of any pattern and a distorted representation of the pattern, which was created by photographing the pattern through a lens provided with the substance to be examined. Instead, a test pattern that is preferably matched to the examination of the substance to be examined is used, and certain texture features are extracted and evaluated. In this way it is possible to draw certain conclusions about the substance to be examined.

Since a smartphone or a tablet computer, for example, can be used as the mobile electronic device, no special hardware, such as an external sensor, is required to carry out the method according to the invention. In addition, the method according to the invention is largely independent of the lighting and existing ambient light with regard to the measurement effect and thus has clear advantages over mobile spectroscopic methods. It is particularly advantageous that the method according to the invention also enables liquid, cream-like or gel-like substances to be examined in a simple manner.

With the method according to the invention, a method for examining substances is thus obtained, with which in particular liquid or gel-like substances can be examined in the simplest and most economical way possible.

The method according to the invention can be carried out in two different variants.

The preferred variant is a relative measurement in which, in addition to the image recorded in step iv), during which the substance to be examined or the mixture is on the camera lens or the protective element, a reference image is recorded which is recorded the substance to be examined or the mixture is not on the camera lens or the protective element (ie the lens is clean when it is picked up). During the extraction in step v) or during the evaluation in step vi), the (preferably calculated and application-specifically optimized) structural features and / or image features of the image recorded when the substance or mixture is present can be combined with the corresponding structural features and / or image features of the reference image be compared. This procedure basically has the advantage that it is relatively independent of the parameters of the camera system used, since these are contained in both images (ie the image recorded when the substance or mixture is present and the reference image), and only the real one through the Substance or the mixture caused difference is considered. In other words, differences caused by the camera system (e.g. faults in the lens) can be filtered out. In addition, this variant can also be used with previously unknown test patterns.

The alternative variant to this is an absolute measurement in which no reference image is recorded. When extracting in step v) or during the evaluation in step vi), the (preferably calculated and application-specifically optimized) structural features and / or image features of the image recorded when the substance or mixture is present can be combined with the corresponding structural features and / or image features of the original Pattern are compared, which is stored for example in a memory of the mobile electronic device. For example, you can determine absolute blurring or distortion and, assuming an ideal camera system, use it as a measuring effect. As the camera quality deteriorates, this measuring effect is increasingly compromised. However, this variant has the advantage that only one measurement has to be carried out, which results in a simpler and more convenient method.

Possibilities for a relative comparison are, for example, the calculation of the pixel-by-pixel deviation from the measurement image to the reference image using the mean square error (MSER) or the calculation of the distance of the histograms, formed on the measurement image and the reference image (e.g. Kullback-Leibler divergence, Chi-Squared divergence ). This process generates a multi-element feature vector for each image.

This procedure can be carried out systematically for all available test patterns.

1. a) Flüssigkeiten,
2. b) eine oder mehrere Flüssigkeiten enthaltenden Gemischen, insbesondere Lösungen, Emulsionen, Suspensionen, flüssige Schäume,
3. c) Gelen und gelartigen Substanzen,
4. d) (flüssigen) Sekreten, beispielsweise Talg (z.B. Hauttalg, Haarfett), Schweiß, Speichel, Milch,
5. e) sowie Mischungen hiervon.

A preferred variant of the method according to the invention is characterized in that the substance to be examined is selected from the group consisting of

1. a) liquids,
2. b) mixtures containing one or more liquids, in particular solutions, emulsions, suspensions, liquid foams,
3. c) gels and gel-like substances,
4. d) (liquid) secretions, e.g. sebum (e.g. sebum, hair fat), sweat, saliva, milk,
5. e) and mixtures thereof.

The mentioned substances have the advantage that they can easily be brought into contact with the camera lens or protective element without prior pretreatment so that part of the substance forms a film (e.g. liquid film, grease film, gel film) on the camera lens or protective element remains. The substances used can be, for example, artificially produced substances (e.g. motor oil, cosmetic creams, lemonade) or natural substances (e.g. blood, sebum, sweat, vegetable oil, tree resin, milk, fruit juice).

• - das in Schritt ii) bereitgestellte Testmuster, und/oder
• - die in Schritt v) extrahierten Texturmerkmale und/oder Bildmerkmale, und/oder
• - die in Schritt vi) durchgeführte Auswertung

auf die bestimmte Substanzart und/oder bestimmte Substanzklasse der zu untersuchenden Substanz abgestimmt ist.According to a further preferred variant of the method according to the invention, the substance to be examined is a substance of a certain type of substance and / or a certain substance class, wherein

• - the test pattern provided in step ii), and / or
• the texture features and / or image features extracted in step v), and / or
• - The evaluation carried out in step vi)

is matched to the specific type of substance and / or specific substance class of the substance to be examined.

• - das in Schritt ii) bereitgestellte Testmuster, und/oder
• - die in Schritt v) extrahierten Texturmerkmale und/oder Bildmerkmale, und/oder
• - die in Schritt vi) durchgeführte Auswertung

auf mindestens eine Eigenschaft der zu untersuchenden Substanz abgestimmt ist, wobei die mindestens eine Eigenschaft ausgewählt ist aus der Gruppe bestehend aus Transparenz der Substanz, Lichtstreuung der Substanz, Lichtbrechung der Substanz, Farbveränderungen der Substanz und deren lokale Verläufe, sowie Kombinationen hiervon.Another preferred variant of the method according to the invention is characterized in that

• - the test pattern provided in step ii), and / or
• the texture features and / or image features extracted in step v), and / or
• - The evaluation carried out in step vi)

is matched to at least one property of the substance to be examined, the at least one property being selected from the group consisting of transparency of the substance, light scattering of the substance, light refraction of the substance, color changes of the substance and their local courses, and combinations thereof.

Another preferred variant of the method according to the invention is characterized in that at least one reference image of the test pattern is additionally recorded with the aid of the camera while the substance or mixture to be examined is not on the camera lens or the protective element. There should preferably be no other substance on the camera lens or the protective element, i.e. the camera lens and / or the protective element should be clean. With this preferred variant, a relative measurement can be carried out. Thus, when selecting the structural features / image features to be extracted and / or when extracting in step v) and / or during the evaluation in step vi), the structural features and / or image features of the image recorded in the presence of the substance or the mixture with the corresponding structural features and / or image features of the reference image are compared. This procedure basically has the advantage that it is relatively independent of the parameters of the camera system used, since these are contained in both images (ie the image recorded when the substance or mixture is present and the reference image), and only the real one through the Substance or the mixture caused difference is considered. In other words, differences caused by the camera system (e.g. faults in the lens) can be filtered out. In addition, this variant can also be used with previously unknown test patterns.

According to a further preferred variant of the method according to the invention, the test pattern (provided in step ii) is a scale-invariant pattern, preferably a fractal, and / or a rotationally symmetrical pattern. The use of a scale invariant pattern is advantageous because the distance between the camera and the test pattern will not always be the same in real smartphone-based applications. The implementation of the method is thus even easier, since it is not specifically based on the distance between the camera and the test pattern must be observed. When using fractals one would have a cyclical basic pattern on a variable scale. The properties of the pattern and its scale plus the properties of the features used are preferably optimized for the (material) properties of the substance to be examined. The use of a rotationally symmetrical pattern has the advantage that when recording the image of the test pattern, one does not have to pay attention to the angle (rotation) of the camera to the test pattern. This also simplifies the method according to the invention.

For example, a Siemens star or a barcode or a QR code can be used as a test pattern.

• - das mobile elektronische Gerät eine Vorrichtung zur Erzeugung von Blitzlicht umfasst und das Testmuster beim Aufnehmen des Bildes in Schritt iv) mithilfe der Vorrichtung zur Erzeugung von Blitzlicht beleuchtet wird, und/oder
• - das Testmuster beim Aufnehmen des Bildes in Schritt iv) durch Umgebungslicht, beispielsweise Tageslicht und/oder Licht aus einer externen Lichtquelle, beleuchtet wird.

Another preferred variant of the method according to the invention is characterized in that

• the mobile electronic device comprises a device for generating flash light and the test pattern is illuminated with the aid of the device for generating flash light when the image is recorded in step iv), and / or
• the test pattern is illuminated by ambient light, for example daylight and / or light from an external light source, when the image is recorded in step iv).

By using a device for generating flash light, the method according to the invention is completely independent of the external lighting conditions. Since smartphones generally have a flashlight device, these are particularly suitable as mobile electronic devices in the method according to the invention.

Another preferred variant of the method according to the invention is characterized in that the test pattern emits at least some areas of electromagnetic radiation in the wavelength range from 1 nm to 100,000 nm, preferably from 10 nm to 10,000 nm, the electromagnetic radiation preferably being selected from the group consisting of radiation from the visible wavelength range, radiation from the near UV range, radiation from the IR range and combinations and mixtures thereof.

The test pattern is preferably a test pattern that fluoresces at least in certain areas.

The test patterns can also consist of (application-specific arranged and spectrally selected) light sources (point or surface light), i.e. not illuminated externally (flash light of the smartphone or ambient light or a combination of both lights), but actively (self-) luminous or fluorescent patterns or parts of the whole Pattern (combination of active and passive patterns / pattern parts). Both the external and the active lighting can be in the wavelength range visible to humans or in the near ultraviolet (UV) or infrared (IR) range. Modern smartphones are already partially equipped with lighting and cameras outside the wavelength range that is visible to humans (e.g. IR). It is pointed out once again that even when non-visible light or light components (external or active) are used, the method according to the invention does not carry out a spectral measurement, but a measurement / evaluation of geometric properties, which, however, can have a better measuring effect in other wavelength ranges.

Another preferred variant of the method according to the invention is characterized in that during the evaluation in step vi) the extracted texture features and / or image features are optimized for the test pattern, with a change in the distance of feature histograms (between measurement of a reference - ie without Substance - and measurement with substance) or (for example in the case of point-like illumination) a point spread function is calculated.

Furthermore, it is preferred that the optimization takes place through an iterative process in which the test pattern and the texture features and / or image features are tested, their combination being checked algorithmically and numerically by selection for their application-specific suitability.

The optimization of the combination of suitable test pattern properties and calculated texture features and / or image features is preferably carried out by an iterative process of testing basic patterns and a wide range of texture features, the combination of which is algorithmically and numerically checked by selection for their application-specific suitability. In concrete terms, this means (1) testing on (non-optimized, general) basic patterns that only meet one property, (2) evaluating a wide range of characteristics for differentiability in the context of the specific measurement task, (3) testing combination patterns that are more as a property, (4) reassessment of the range of successive specific features for differentiability in the context of the specific measurement task. The iteration from (1) to (4) is preferably run through at least once, particularly preferably several times.

According to a further preferred variant of the method according to the invention, the at least one auxiliary is selected from the group consisting of solvents, solubilizers, consistency regulators, gel formers, emulsifiers, wetting agents, fluxes, dyes, color pigments, fluorescent agents, and mixtures thereof. The use of auxiliaries such as solvents, solubilizers, consistency regulators, gel formers, emulsifiers, wetting agents and fluxes can, for example, make it easier to bring the substance to be examined into contact with the camera lens or the protective element. By means of wetting agents and / or fluxing agents, a more uniform local distribution of the substance to be examined or the reduction of the surface tension can be achieved, whereby the measurement can be significantly improved in each case. The mentioned auxiliaries can also change the transparency or the refractive index in such a way that the image can be recorded better. Dyes, color pigments and fluorescent agents can be used to change the transparency, the contrast or the light scattering in a targeted manner so that the image can be recorded better.

As a further preferred variant, there is the possibility within the proposed invention of using the combination of pattern and feature processing both on approximately homogeneous properties of the substance to be examined (e.g. oil / fat) and on substances that occur very selectively due to high surface tension (e.g. Waterdrop). For example, the geometric shape (especially size) of a single drop or the geometric arrangement of many small drops on the camera lens or the protective element can change depending on the surface tension and this in turn can change depending on the chemical properties of the substance to be examined. Illustrative, but not restrictive, the lotus effect and crystallization (not necessarily at temperatures below 0 ° C) are mentioned here. In this case, patterns and features can be optimized additionally or exclusively to these properties.

• - die Qualität der zu untersuchenden Substanz bestimmt wird, und/oder
• - mindestens eine Eigenschaft der zu untersuchenden Substanz bestimmt wird, und/oder
• - die Zugehörigkeit der zu untersuchenden Substanz zu einer bestimmten Substanzart und/oder Substanzklasse bestimmt wird, und/oder
• - die Echtheit der zu untersuchenden Substanz bestimmt wird, und/oder
• - die chemische Zusammensetzung der zu untersuchenden Substanz bestimmt wird.

Another preferred variant of the method according to the invention is characterized in that during or through the evaluation in step vi)

• - the quality of the substance to be examined is determined, and / or
• - At least one property of the substance to be examined is determined, and / or
• - the affiliation of the substance to be examined to a certain substance type and / or substance class is determined, and / or
• - The authenticity of the substance to be examined is determined, and / or
• - the chemical composition of the substance to be examined is determined.

The mobile electronic device is preferably a smartphone or a tablet computer. In this way, a particularly simple implementation of the method according to the invention is made possible, since these are very easy to operate, very familiar and easily accessible devices for most users.

There are a number of applications for the process according to the invention, which are mentioned here but are not intended to have a restrictive effect.

For example, an application in which you pull the oil dipstick out of the car engine and tap it with the smartphone camera (possibly with a rag) and see whether the engine needs an oil change or not. The properties of the oil change with its mileage, and it may even have contained abrasion particles. This should all lead to diffraction.

Another application is the evaluation of essential chemical components on the surface of skin and hair (e.g. sebum, especially fat, moisture) and the derivation of personalized recommendations for cosmetic care products and / or recommendations for action for handling skin and hair (e.g. frequency of the Washing).

In the food sector, this can be used to determine potentially essential properties (e.g. sugar).

• - Anwendungen, die primär auf eine Qualitätsbewertung der zu untersuchenden Substanz abzielen, entweder quantitativ (kontinuierliche Maßzahl) oder qualitativ (Einteilung in Klassen/Gruppen);
• - Anwendungen, die primär der Unterscheidung von Substanzen bzw. Objekten anhand ihrer chemischen Eigenschaften abzielen, unabhängig davon, ob diese Unterscheidung ein Qualitätsmerkmal darstellt;
• - Anwendungen, die primär auf das Erkennen der Echtheit bzw. von Fälschungen abzielen, entweder anhand der natürlich auftretenden chemischen Zusammensetzung der Probe oder durch gezieltes Einbringen zusätzlicher Substanzen, die als eine Art künstliche Signatur den o.g. Messeffekt unterstützen. Während in den vorgenannten zwei Anwendungen davon auszugehen ist, dass die Definition/Beschreibung der Klassen a-priori bekannt ist (Binär- oder Multiklassifikation), besteht hier die Aufgabe einer Abgrenzung einer bekannten Klasse gegen einen unbekannten Rest (unbekannte Fälschungen, 1-vs.-Rest-Klassifikation).

On a higher level of abstraction, categories of applications can be named:

• - Applications that primarily aim at a quality assessment of the substance to be examined, either quantitatively (continuous measure) or qualitatively (division into classes / groups);
• - Applications that primarily aim to differentiate between substances or objects on the basis of their chemical properties, regardless of whether this differentiation represents a quality feature;
• - Applications that are primarily aimed at recognizing authenticity or forgeries, either on the basis of the naturally occurring chemical composition of the sample or through the targeted introduction of additional ones Substances that support the above-mentioned measuring effect as a kind of artificial signature. While in the above two applications it can be assumed that the definition / description of the classes is known a priori (binary or multiclassification), the task here is to delimit a known class from an unknown remainder (unknown forgeries, 1-vs. -Rest classification).

The present invention also relates to a computer program product for use in the method according to the invention, wherein the computer program product can be loaded directly into an internal data memory of the mobile electronic device provided in step i) and comprises software sections with which at least steps v) and vi) (of the invention Method), preferably steps iv), v) and vi) (of the method according to the invention), are carried out when the computer program is running on the mobile electronic device.

The present invention further relates to a mobile electronic device on which a computer program product according to the invention is installed, the mobile electronic device preferably being a smartphone or a tablet computer. This mobile electronic device is then the mobile electronic device used in the method.

The present invention is explained in more detail with reference to the following figures and examples, without restricting the invention to the parameters specifically shown.

Examples of test patterns

1. 1) im Hinblick auf gerade, in unterschiedlichen Winkeln verlaufende scharfe (kontrastreiche) Kanten, die durch den Messeffekt verändert werden;
2. 2) im Hinblick auf Rotationsinvarianz, die eine bestimmte Ausrichtung zwischen Messmuster und Kamerasystem in weiten Grenzen überflüssig macht;
3. 3) im Hinblick auf unterschiedliche Skalen, an denen man den Messeffekt besonders gut erkennen kann, z.B. durch Verschwimmen / Verschmieren der schwarzen und weißen Sektoren abhängig von deren Abstand zueinander entlang deren Verlaufes vom Mittelpunkt zum Rand;
4. 4) im Hinblick auf unterschiedliche Skalen, die eine Größeninvarianz herbeiführen und somit einen festen Abstand zwischen Messmuster und Kamerasystem in weiten Grenzen überflüssig macht.

Basically, the group of fractals are very suitable test patterns. A particularly suitable and clear exemplary embodiment for a test pattern is the so-called Siemens star, which is exemplified in 1 is shown. This combines essential of the above-mentioned preferred properties of the test pattern to be used, among others

1. 1) with regard to straight, sharp (high-contrast) edges running at different angles, which are changed by the measuring effect;
2. 2) with regard to rotational invariance, which makes a specific alignment between the measurement pattern and the camera system superfluous to a large extent;
3. 3) with regard to different scales, on which the measuring effect can be seen particularly well, for example by blurring / smearing the black and white sectors depending on their distance from one another along their course from the center to the edge;
4. 4) with regard to different scales, which cause size invariance and thus make a fixed distance between the measurement pattern and the camera system superfluous within wide limits.

Another suitable exemplary embodiment for a possible test pattern are bar codes or, even better, QR codes. These often, especially in the case of QR codes, provide at least a limited feature of scale and size invariance (due to boxes of different sizes) and a certain rotational invariance with a fine structure and resolution. With regard to the application of the described invention in the consumer sector, it should be emphasized as particularly advantageous that such codes are usually already present on packaging or supplementary sheets / descriptions, and it is not necessary to change these when using the described invention. An explicit provision of a dedicated measurement pattern is therefore not required. In this exemplary embodiment, the optimization described is limited to application-specific image filters / texture features. As an extension, however, such codes can be designed in such a way that, on the one hand, they fulfill their original purpose of providing easily machine-readable information, and, on the other hand, support their use as measurement patterns within the scope of the proposed invention, e.g. through redundant coding and the resulting higher filigree resolution.

Another suitable embodiment is the use of text, preferably in different font sizes. An application-specific optimization can be achieved through the recognition and targeted evaluation of certain characters (letters and numbers) and their disturbed mapping, caused by the measurement effect. The same applies here in connection with QR codes and barcodes, in particular the easy availability for applications in the consumer sector.

Even if in some of the exemplary embodiments, in particular the last-mentioned, the optimized design of the actual measurement pattern for the application is not given, the assessment of the basic usability in the context of a specific application is ultimately also part of the optimization of the measurement pattern and thus part of the proposed invention.

Exemplary embodiment: engine oil measurement scenario

In this exemplary embodiment, the method according to the invention for examining (in engine oil located in an engine of a motor vehicle and already used there) in order to draw a conclusion about the quality of the engine oil and to be able to decide whether a change of the engine oil is already necessary or not.

For this purpose, the user initially provides a smartphone or a tablet computer which has a camera with a camera lens. The user pulls the dipstick out of the engine and then rubs oil off the dipstick with a paper tissue or other wiping element and then on the camera lens of the smartphone or tablet computer. This creates a film of engine oil on the camera lens of the smartphone or tablet computer without affecting the smartphone or tablet computer.

Subsequently (i.e. while the film of the engine oil is on the camera lens) the user photographs the test pattern, which is provided, for example, as a print on the oil bottle or on the paper tag that is typically attached to the dipstick when changing the oil, or it was printed out from the Internet . With a numerical processing pipeline stored in the smartphone or tablet computer, consisting of an extraction of texture features from the recorded image, which differ from the corresponding texture features of the original test pattern, and an evaluation in the form of a mathematical detection model, the oil film is generated using a three-class system according to a traffic light system rated, green = oil in normal condition, yellow = change recommended, red = change necessary.

In order to reach the numerical processing pipeline, a representative sample of various motor oils available on the market in various aging stages can be collected in advance (e.g. by the developer of the detection system) within a measurement campaign. In addition, an expert in engine technology can evaluate the recorded reference samples using laboratory technology in order to create the "basic truth" for the detection system, on the basis of which each sample can be classified into the above-mentioned class system.

1. 1) Bildoperation zur Bewertung der globalen Bildentropie,
2. 2) Bildoperation zur Bewertung der globalen Weichzeichnung (eng. Bluriness) im Bild,
3. 3) Bildoperation zur Bewertung des Mittelwertes, der Standardabweichung und des Kontrastes eines lokalen Bildausschnittes,
4. 4) Bildoperation zur Bewertung der Orientierung und Stärke von Bildkanten (schwarz-weiß Übergänge) mittels, z.B. Gaborfilter,
5. 5) Bildoperation zur Errechnung einer 2D Fourier Transformation zur Bewertung von periodischen Mustern im Bild.

A number of image filter operations are available to the developer for capturing features, e.g.

1. 1) image operation to evaluate the global image entropy,
2. 2) image operation to evaluate the global bluriness in the image,
3. 3) Image operation to evaluate the mean value, standard deviation and contrast of a local image section,
4. 4) Image operation to evaluate the orientation and strength of image edges (black-white transitions) by means of, e.g. Gabor filters,
5. 5) Image operation to calculate a 2D Fourier transformation for evaluating periodic patterns in the image.

1. 1) als Absolutwerte,
2. 2) als Relativwerte zu einer vor-Ort Aufnahme des Testmusters mit sauberer Linse,
3. 3) als Relativwerte zu einer intern hinterlegten Aufnahme des Testmusters bei sauberer Linse

Each image in the problem-specific database is filtered with all or some of the image operators mentioned. The image features obtained can be used for the ultimate feature extraction

1. 1) as absolute values,
2. 2) as relative values to an on-site recording of the test sample with a clean lens,
3. 3) as relative values to an internally stored recording of the test pattern with a clean lens

Find use. Possibilities for a relative comparison are, for example, the calculation of the pixel-by-pixel deviation from the measurement image to the reference image using the mean square error (MSER) or the calculation of the distance of the histograms, formed on the measurement image and the reference image (e.g. Kullback-Leibler divergence, Chi-Squared divergence ).

This process generates a multi-element feature vector for each image. Using a learning algorithm, the parameters of a mathematical model are then adapted in such a way that the error in the prediction of the class based on the feature vector is minimal. Several methods, e.g. Artificial Neural Networks (ANN), Support Vector Machines (SVM) or Partial Least Squares (PLS) models are tested for their suitability and selected based on a performance criterion.

This procedure can be carried out systematically for all available test patterns.

To find the optimal combination of test pattern and feature extraction, a problem-specific ordering of the importance of the features is carried out on the basis of the internal weighting of the features in the mathematical model. The model creation is then systematically repeated, in which successive features are added to the input of the model in order of their importance. This process is carried out until an equivalent performance compared to the system based on all features is achieved.

The application-specific combination of feature and results from this consideration Test pattern for the task, e.g. the classification of the need for an oil change.

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• WO 2018/115346 [0004]

Claims (17)

Method for the investigation of substances in which i) a mobile electronic device is provided which has a camera with a camera lens, ii) a test pattern is provided, iii) a substance to be investigated or a mixture of a substance to be investigated and at least one auxiliary substance is brought into contact with the camera lens or with a protective element arranged on or in front of the camera lens, so that a film of the substance to be investigated or the mixture is on the Camera lens or protective element, iv) at least one image of the test pattern is recorded with the aid of the camera while the film of the substance to be examined or the mixture is on the camera lens or the protective element, the recorded image of the test pattern showing an at least partially distorted representation of the test pattern, with at least some of the texture features and / or image features of the at least partially distorted representation of the test pattern differ from the corresponding texture features and / or image features of the original test pattern, v) at least some of the differing texture features and / or image features are extracted from the recorded image of the test pattern, and vi) the extracted texture features and / or image features are evaluated. Method according to the preceding claim, characterized in that the substance to be examined is selected from the group consisting of a) liquids, b) mixtures containing one or more liquids, in particular solutions, emulsions, suspensions, liquid foams, c) gels and gel-like substances , d) secretions, e) and mixtures thereof. Method according to one of the preceding claims, characterized in that - the test pattern provided in step ii), and / or - the texture features and / or image features extracted in step v), and / or - the evaluation carried out in step vi) on the investigation matched to the substance to be examined. Method according to one of the preceding claims, characterized in that the substance to be examined is a substance of a certain substance type and / or certain substance class, wherein - the test pattern provided in step ii), and / or - the texture features extracted in step v) and / or image features, and / or the evaluation carried out in step vi) are matched to the specific type of substance and / or specific substance class of the substance to be examined. Method according to one of the preceding claims, characterized in that - the test pattern provided in step ii), and / or - the texture features and / or image features extracted in step v), and / or - the evaluation carried out in step vi) for at least one Property of the substance to be examined are matched, the at least one property being selected from the group consisting of transparency of the substance, light scattering of the substance, light refraction of the substance, color changes of the substance and their local courses, and combinations thereof. Method according to one of the preceding claims, characterized in that in addition at least one reference image of the test pattern is recorded with the aid of the camera while the substance or mixture to be examined is not on the camera lens or the protective element. Method according to one of the preceding claims, characterized in that the test pattern is a scale-invariant pattern, preferably a fractal, and / or a rotationally symmetrical pattern. Method according to one of the preceding claims, characterized in that - the mobile electronic device comprises a device for generating flash light and the test pattern is illuminated when taking the image in step iv) with the aid of the device for generating flash light, and / or - the test pattern when taking the image in step iv) is illuminated by ambient light, for example daylight and / or light from an external light source. Method according to one of the preceding claims, characterized in that the test pattern emits at least regionally electromagnetic radiation in the wavelength range from 1 nm to 100,000 nm, preferably from 10 nm to 10,000 nm, the electromagnetic radiation preferably being selected from the group consisting of radiation from the visible wavelength range, radiation from the near UV range, radiation from the IR range and combinations and mixtures thereof. Method according to one of the preceding claims, characterized in that the test pattern is a test pattern that fluoresces at least in certain areas. Method according to one of the preceding claims, characterized in that during the evaluation in step vi) the extracted texture features and / or image features are optimized for the test pattern, a change in the distance of feature histograms or a point spreading function being preferably calculated. Method according to the preceding claim, characterized in that the optimization takes place by an iterative process in which the test pattern and the texture features and / or image features are tested, their combination being checked algorithmically and numerically by selection for their application-specific suitability. Method according to one of the preceding claims, characterized in that the at least one auxiliary is selected from the group consisting of solvents, solubilizers, consistency regulators, gel formers, emulsifiers, wetting agents, fluxes, dyes, color pigments, fluorescent agents, and mixtures thereof. Method according to one of the preceding claims, characterized in that in the evaluation in step vi) - the quality of the substance to be examined is determined, and / or - at least one property of the substance to be examined is determined, and / or - the affiliation to substance to be examined is determined for a certain type of substance and / or substance class, and / or - the authenticity of the substance to be examined is determined, and / or - the chemical composition of the substance to be examined is determined. Method according to one of the preceding claims, characterized in that the mobile electronic device is a smartphone or a tablet computer. Computer program product for use in a method according to one of the preceding claims, wherein the computer program product can be loaded directly into an internal data memory of the mobile electronic device provided in step i) and comprises software sections with which at least steps v) and vi), preferably steps iv), v) and vi), are executed when the computer program is running on the mobile electronic device. Mobile electronic device on which a computer program product according to Claim 16 is installed, wherein the mobile electronic device is preferably a smartphone or a tablet computer.

Images